New Practical Synthesis of Indazoles via Condensation of o-Fluorobenzaldehydes and Their O-Methyloximes with Hydrazine Kirill Lukin,* Margaret C. Hsu, Dilinie Fernando, and M. Robert Leanna GPRD Process Research and DeVelopment, Abbott Laboratories, North Chicago, Illinois 60064
[email protected] ReceiVed July 3, 2006
The reaction of o-fluorobenzaldehydes and their O-methyloximes with hydrazine has been developed as a new practical synthesis of indazoles. Utilization of the methyloxime derivatives of benzaldehydes (in the form of the major E-isomers) in this condensation effectively eliminated a competitive Wolf-Kishner reduction to fluorotoluenes, which was observed in the direct preparations of indazoles from aldehydes. Reaction of Z-isomers of methyloximes with hydrazine resulted in the formation of 3-aminoindazoles via a benzonitrile intermediate.
Introduction Investigation of biologically active compounds possessing the indazole (1) heterocyclic core has resulted in the discovery of potent HIV protease inhibitors, serotonin receptor antagonists, aldol reductase inhibitors, and acetylcholinesterase inhibitors.1,2 Recently, another indazole derivative, ABT-102 (2), has been identified as a potent vanilloid receptor (VR1) antagonist. Compound 2 is currently undergoing advanced clinical development for the treatment of chronic pain.3
The current synthesis of 2 utilizes 4-haloindazoles 3 or 4 as starting materials. Development of a reliable and efficient preparation of these indazoles was required to provide access to large quantities of bulk drug for the studies. Inspection of (1) Rodgers, J. D.; Johnson, B. L.; Wang, H.; Greenberg, R. A.; EricksonViitanen, S.; Klabe, R. M.; Cordova, B. C.; Reyner, M. M.; Lam, G. N.; Chang, C.-H. Bioorg. Med. Chem. Lett. 1996, 6, 2919-2924. (2) See, for example: Brase, S.; Gil, C.; Knepper, K. Bioorg. Med. Chem. 2002, 10, 2415-2437. (3) Gomtsyan, A.; Bayburt, E. K.; Koenig, J. R.; Lee, C.-H. U.S. Patent Application 2004254188, Dec 16, 2004.
the literature4 revealed that the indazoles substituted on the sixmembered ring (including 3 and 4) were generally prepared via diazotation of the corresponding toluidines 5 (eq 1)5,6 or the
nitrozation of their N-acetyl derivatives 7 (Jacobsen modification, eq 2).7,8 Unfortunately, these methods could not be considered practical for multikilogram preparations.9,10 Other innovative approaches for indazoles synthesis have been reported, but they either gave mixtures of isomeric products11 or were only limited to the synthesis of indazoles substituted at the 1-N or 3-C positions.4,12 However, among the latter methods, (4) For reviews see: (a) Elguero, J. In ComprehensiVe Heterocyclic Chemistry; Katrizky, A. R., Rees, C. W., Eds.; Pergamon Press: New York, 1984; Vol. 5, p 167. (b) Behr, L. C.; Fusco, R.; Jarobe, C. H. In Pyrazoles, Pyrazolines, Pyrazolidines, Indazoles and Condensed Rings; Wiley, R. H., Ed.; Wiley Int.: New York, 1969; p 28. (5) Ruechardt, C.; Hassmann, V. Liebigs Ann. Chem. 1980, 6, 908927. (6) For other examples of indazoles synthesis via diazonium salt see: Schumann, P.; Collot, V.; Hommet, Y.; Gsell, W.; Dauphin, F.; Sopkova, J.; MacKenzie, E. T.; Duval, D.; Boulouard, M.; Rault, S. Bioorg. Med. Chem. Lett. 2001, 11, 1153-1156. (7) (a) Tono-oka, S.; Tone, Y.; Marques, V. E.; Cooney, D.; Sekikawa, I.; Azuma, I. Bull. Chem. Soc. Jpn. 1985, 58, 309-315. (b) Shoda, M.; Kuriyama, H. U.S. Patent Application 2003070686, Aug 28, 2003. (8) (a) Jacobsen, P.; Huber, L. Chem. Ber. 1908, 41, 660. (b) Ruchardt, C.; Hassmann, V. Liebigs Ann. Chem. 1980, 908-927. 10.1021/jo0613784 CCC: $33.50 © 2006 American Chemical Society
8166
J. Org. Chem. 2006, 71, 8166-8172
Published on Web 09/15/2006
Synthesis of Indazoles from o-Fluorobenzaldehydes
Surprisingly though, we could not find any references of similar condensations involving o-fluorobenzaldehydessstarting materials required for the preparation of indazoles unsubstituted at the 3-position. The results of our investigation of the latter reaction, which led to the development of a new practical synthesis of indazoles, are provided below. Results and Discussion a condensation of ortho-substituted benzaldehydes with hydrazines (eq 3) looked particularly straightforward and attractive
for the adaptation for the synthesis of indazoles 3 and 4. The condensation outlined in eq 3, utilizing nitro as a leaving group (Z ) NO2) was first reported more than a century ago.13 It was demonstrated that several substituted hydrazones (R * H), which were initially formed in the reactions of o-nitrobenzaldehydes with the corresponding hydrazines, could be further converted into the indazoles in the presence of a base under very forcing conditions. Later, this method was adapted by employing the corresponding o-fluoro- and o-mesyloxysubstituted14 benzocarbonyl compounds. Several groups reported that various 3-substituted indazoles 9 (R * H, eq 4) could be
efficiently prepared via condensations of the corresponding esters of 2-fluorobenzoic acid (10a, R ) OMe), fluoroaceto, and fluorobenzophenones (10a, R ) Me, Ar), as well as 2-fluorobenzonitriles 10b, with hydrazine.15-18 (9) The diazotation method generally requires the isolation of a potentially explosive diazonium salt intermediate. The formation and cyclization of the diazonium salt proceed simultaneously under the reaction conditions only in the presence of additional highly electron-withdrawing groups (e.g., nitro): (a) Bartsch, R. A.; Tang, I.-W. J. Heterocycl. Chem. 1984, 21, 1063. (b) Porter, H. D.; Peterson, W. D. Org. Synth. 1955, III (Collective Vol.), 660. (10) Even the most advanced modification of the Jacobsen reaction remains relatively impractical, as it requires the addition of 0.1 equiv of expensive and toxic crown ether catalyst, employs toxic chloroform as a preferred solvent, and adds a deprotection step to recover the indazole: (a) Beadle, J. R.; Korzeniowski, S. H.; Rosenberg, D. E.; Garcia-Slanga, B. J.; Gokel, G. W. J. Org. Chem. 1984, 49, 1594-1603. (b) Sun, J.-H.; Teleha, C. A.; Yan, J.-S.; Rodgers, J. D.; Nugiel D. A. J. Org. Chem. 1997, 62, 5627-5629. (11) Bromoindazole 3 was obtained as one of the products resulting from the condensation of bromobenzyne with (trimethylsilyl)diazomethane: Shoji, Y.; Hari, Y.; Aoyama, T. Tetrahedron Lett. 2004, 45, 1769-1771. (12) Song, J. J.; Yee, N. K. Tetrahedron Lett. 2001, 42, 2937-2940. (13) (a) Meyer, V. Chem. Ber. 1889, 22, 319. (b) Reich, S.; Gaigalian, G. Chem. Ber. 1913, 46, 2380-2387. (14) Caron, S.; Vazquez, E. Synthesis 1999, 4, 588-592. (15) Cui, J. J.; Araldi, G.-L.; Reiner, J. E.; Reddy, K. M.; Kemp, S. J.; Ho, J. Z.; Siev, D. V.; Mamedova, L.; Gibson, T. S.; Gaudette, J. A.; Minami, N. K.; Anderson, S.; M.; Bradbury, A. E.; Nolan, T. G.; Semple, J. E. Bioorg. Med. Chem. Lett. 2002, 12, 2925-2930. (16) Henke, B. R.; Aquino, C. J.; Birkemo, L. S.; Croom, D. K.; Dougherty, R. W.; Ervin, G. N.; Grizzle, M. K.; Hirst, G. C.; James, M. K.; Johnson, M. F.; Queen, K. L.; Sherrill, R. G.; Sugg, E. E.; Suh, E. M.; Szewczyk, J. W.; Unwalla, R. J.; Yingling, J.; Willson, T. M. J. Med. Chem. 1997, 40, 2706-2725.
1. Synthesis of 4-Haloindazoles. Our initial investigation into the synthesis of 4-bromo and 4-chloroindazoles (3, 4) via a condensation of the corresponding 6-halo-2-fluorobenzaldehydes 11 and 12 with hydrazine gave fairly promising results. For example, when the reaction of aldehyde 11 with hydrazine hydrate (3 equiv) in the presence of sodium bicarbonate was monitored by HPLC, a fast and quantitative formation of hydrazone intermediate 13 was initially observed (eq 5). Further
heating of this mixture to 90-100 °C for 5 h resulted in conversion of 13 into an approximately 2:3 mixture of the desired indazole 3 and an unknown side product. The side product was subsequently identified as bromofluorotoluene 14,19 apparently resulting from the Wolf-Kishner type reduction of the intermediate hydrazone 13 (eq 5). Similar results were observed with 6-chloro analogue 12. As it was generally presumed that the formation of indazoles from ortho-substituted hydrazones proceeded as an intramolecular cyclization (eq 3), we decided to further optimize the reaction conditions for this cyclization using the isolated pure hydrazone 13. First, we found that a careful selection of a base was critical for minimizing the Wolf-Kishner pathway. Thus, in the presence of bases stronger than sodium bicarbonate (e.g., potassium carbonate, triethylamine) fast and quantitative reduction of hydrazone 13 to toluene 14 was observed. The formation of indazole 3 was only observed in the presence of weak bases (e.g., sodium bicarbonate, pyridine). However, the cyclization of the isolated hydrazone 13 proceeded remarkably slower, compared to the described above reaction of the same hydrazone, when it was formed in situ. Even after 72 h at 100 °C, some of the starting hydrazone remained in the mixture. Moreover, the yield of indazole 3 obtained in this experiment was lower than 10%, due to the formation of toluene 14 and other numerous side products. These data suggested that the (17) (a) Shutske, G. M.; Allen, R. C.; Forsch, M. F.; Setescak, L. L.; Wilker, J. C. J. Med. Chem. 1983, 26, 1307-1311. (b) Dehmlow, H.; Aebi, J. D.; Jolidon, S.; Ji, Y.-H.; Mark, E. M.; Himber, J.; Morand, O. H. J. Med. Chem. 2003, 46, 3354-3370. (18) Witherington, J.; Bordas, V.; Gaiba, A.; Naylor, A.; Rawlings, A. D.; Slingsby, B. P.; Smith, D. G.; Takle, A. K.; Ward, R. W. Bioorg. Med. Chem. Lett. 2003, 13, 3059-3062. (19) Dewar, M. J. S.; Grisdale, P. J. J. Org. Chem. 1963, 28, 17591762.
J. Org. Chem, Vol. 71, No. 21, 2006 8167
Lukin et al. SCHEME 1
formation of indazoles in reactions of o-fluorobenzaldehydes with hydrazine under the investigated set of conditions did not proceed via an intramolecular cyclization of hydrazone 13.20 In search of an explanation for the above results, we found that the indazole formation from 11 could be facilitated in the presence of excess hydrazine. Indeed, when the isolated hydrazone 13 was added to hydrazine hydrate, which was used as a solvent, heating the solution to 100 °C for only 1 h resulted in consumption of the starting material. Based on these data we rationalized that the mechanism for the conversion of the o-fluorophenylhydrazones into indazoles in the presence of excess hydrazine proceeded by substitution of the aryl fluoride with another molecule of hydrazine, followed by the cyclization of 13b as outlined in Scheme 1. According to this mechanism, the hydrazine molecule, which was initially incorporated into the hydrazone fragment of 13, would be later eliminated as a result of a subsequent cyclization step via aminal 13c (Scheme 1). To support this proposed mechanism we prepared benzyl-substituted hydrazone 16 and subjected it to the cyclization conditions in hydrazine hydrate (eq 6). As we expected, nonbenzylated indazole 3 was formed
as a major product (the 10:1 ratio of indazole 3 and benzylindazole 17 was determined by HPLC methodology). Importantly, only traces of toluene side product were observed in the reaction mixture (in comparison to the cyclization of 13 which gave 3040% of reduction product), suggesting that hydrazone 13 was not a major intermediate in this reaction and that the indazole formation proceeded in an intramolecular fashion, after the fluoride substitution in 16 (Scheme 1). As an outcome of the above mechanistic rationalization, we were able to prepare 4-haloindazoles 3 and 4 simply by refluxing the corresponding aldehydes in hydrazine hydrate. Upon the completion of the reaction, the product typically precipitated out during cooling of the crude reaction mixture and was conveniently isolated through filtration. The yields of indazoles 3 and 4 obtained in this reaction were satisfactory (50-60%), (20) It is possible that intramolecular cyclization of hydrazones with the formation of indazoles still could be accomplished at higher temperatures. In the related study of the intramolecular cyclizations of 2-mesyloxyacetophenones hydrazones (ref 18, eq 4), the indazole formation was observed at temperatures above 130 °C. However, a DSC scan of hydrazone 13 indicated that it underwent highly exothermic decomposition at 140-150 °C, prompting us to set 110 °C as the upper limit for all our experiments.
8168 J. Org. Chem., Vol. 71, No. 21, 2006
however, still negatively affected by the formation of toluenes 14 and 15, as side products. Upon further optimization of the reaction parameters, we have found that the extent of the reduction could be attenuated by conducting the reaction in etheral solvents such as THF, DME, or dioxane and employing anhydrous hydrazine as a cosolvent. Under these optimized conditions (1:1 THF-98% hydrazine, 70 °C, 15 h) the isolated yields of indazoles 3 and 4 were increased to 80-85%, respectively. 2. Condensation of Other o-Fluorobenzaldehydes with Hydrazine. In an attempt to determine the generality of this practical approach, the condensations of several substituted o-fluorobenzaldehydes (18-24) were evaluated using the conditions found optimal for aldehydes 11 and 12. The selection of the o-fluorobenzaldehydes was designed to evaluate both the impact of the substitution pattern and the electronic effect of substituents (eq 7). The results of this study are summarized in
Table 1. We were pleased to find that the desired indazoles could be obtained in good yields in most of the cases studied. At the same time, the data in Table 1 show that the cyclizations of unsubstituted fluorobenzaldehyde 18 and 5-bromo-2-fluorobenzaldehyde (20) gave reduced yields of the corresponding indazoles, whereas 5-methoxy-substituted aldehyde 23 gave only the corresponding toluene reduction product. The lower efficiency of the condensation reaction in the latter examples could be rationalized by considering the effect of substituents on the rates of the two competitive processessnucleophilic substitution of the fluoride leading to indazole formation and the hydrazone intermediate reduction leading to the toluene. For example, introduction of the electron-donating methoxy substituent in para or ortho positions relative to the carbonyl group (compounds 22, 24) would presumably decelerate the hydrazone reduction, while having little effect on the fluoride substitution. At the same, introduction of the methoxy substituent in the para position relative to the fluoride (compound 23) would deactivate the fluoride substitution, with little effect on the hydrazone reduction rate. Accordingly, the condensation of aldehyde 24 resulted in a high yield of 4-methoxyindazole (31), whereas the analogous reaction of aldehyde 23 gave the reduction product only. Although in all studied cases the toluene side products could be conveniently removed from the indazoles by a heptane wash, or simply during the product drying under vacuum, we still sought better control over the reduction pathway to achieve higher yields of indazoles regardless of the substitution. 3. Condensation of 2-Fluorobenzaldehyde O-Methyloximes with Hydrazine. The condensation of benzylhydrazone 16 with hydrazine (eq 6) provided important clues for further method development. As noted above, this reaction resulted in replacement of benzylhydrazine and gave nearly exclusively indazole unsubstituted at the 1-N position. These findings allowed us to propose that not only various substituted hydrazones (e.g., 16) but other appropriately protected aldehydes could be utilized in the condensations with hydrazine for indazoles preparation. Moreover, elimination of the hydrazone intermediate from this new process would avoid the Wolf-Kishner pathway leading to the toluene side products. We thought that easily accessible oximes could be good surrogates to investigate this approach.
Synthesis of Indazoles from o-Fluorobenzaldehydes TABLE 1. Indazole Formation from Benzaldehydes 11, 12, 18-24
the standard conditions using the hydrazone intermediate (Table 1, entry 3). Adding to the convenience of this procedure, no isolation of the methyloxime intermediate was required before the indazole formation step. The reaction with hydrazine was conducted simply after filtering off the inorganics from the crude oxime solution (vide infra). In a similar manner, the benzaldehydes referenced in Table 1, as well as 2,6-difluorobenzaldehyde (39), were converted into the corresponding methyloximes and, subsequently, indazoles in high yields, as summarized in Table 2. The only exception to this generality was the methyloxime of highly deactivated 5-methoxy-substituted aldehyde 23, which upon extended reaction time gave only 5% yield of 30. Surprisingly, 6-bromo- and 6-chloro-substituted benzaldehydes 11 and 12, which gave the highest yields of indazoles in the cyclizations via hydrazone intermediate (Table 1, entries 1 and 2), behaved unusually in the cyclization via their requisite methyloxime. It was found that in addition to the corresponding indazoles 3 and 4, other more polar products (12-14% by HPLC peak area vs 3 or 4) were formed in these reactions (eq 10).
a
Reaction yield, as determined by HPLC assay. b Additionally, toluene 15 was formed in 13% yield (HPLC assay). c Additionally, 2-fluorotoluene was formed in 45% yield (HPLC assay).
However, it has been reported that unprotected oximes of substituted 2-fluorobenzaldehydes could undergo a self-cyclization into isoxazoles, as outlined in eq 8.21
The isoxazole side product formation was obviated by using O-methyloximes 32 instead of oximes. Compounds 32 could be simply prepared in quantitative yields by reacting the equivalent amounts of fluorobenzaldehydes and methyloxime hydrochloride in the presence of potassium carbonate (eq 9). We were pleased to find that the reaction of 2-fluorobenzaldehyde O-methyloxime prepared in this manner with hydrazine cleanly gave indazole, which was isolated in 80% yield (eq 9, X ) H) representing a nearly 3-fold yield improvement over (21) Strupczewski, J. T.; Allen, R. C.; Gardner, B. A.; Schmid, B. L.; Stache, U.; Glamkowski, E. J.; Jones, M. C.; Ellis, D. B.; Huger, F. P.; Dunn, R. W. J. Med. Chem. 1985, 28, 761-769.
These side products were identified as 3-aminoindazoles 34a and 34b,22 respectively. It was also observed that formation of 34 could be related to the increased amounts of the minor Z-isomers of oximes 35 or 36 formed during the oximation of aldehydes 11 and 12, respectively (see Table 2).23 To confirm that 3-aminoindazoles were indeed formed from the corresponding Z-oxime isomers, we attempted the isolation of individual compounds 35a and 35b. Although the separation of these isomers via chromatography was not achievable, it was possible to isolate the major E-form 35b ( Cl > MeO > F > H, see Table 2). The same substituents did not significantly affect the ratio of oxime isomers, if they were placed in other positions in the molecule. Accordingly, we did not observe the formation of any side products with an HLPC peak area higher than 5% in the condensations of the latter oximes (see Table 2).24 A plausible mechanism for the observed formation of 3-aminoindazoles in the condensations of Z-methyloximes with hydrazine is outlined in eq 11.
This mechanism is based on the reported literature precedents.14,17 It was also demonstrated that O-methyloximes of benzaldehydes, possessing electron-withdrawing groups, could be converted into the corresponding nitriles in the presence of bases as strong as aqueous sodium hydroxide.25 The selective benzonitrile formation from the Z-isomer of oxime under the reaction conditions probably represented a kinetic effect. Relatively faster elimination of methanol from the Z-isomer could be explained by the preferred orbital orientation in the elimination transitions state of this isomer, in line with the general stereochemical considerations for the reactions proceeding via an E2 mechanism.26 Further support of the mechanism outlined in eq 11 was obtained during monitoring the condensation of oximes 36 with hydrazine by HPLC. While no intermediates were observed in the cyclization of 36b into 4, the formation of aminoindazole 34b clearly proceeded via an intermediate (40% peak area vs 34b after 1 h; 10% after 2 h; 95 : 5) was obtained by crystallization of the mixture from pentane at 0 °C: 1H NMR (CDCl , δ, ppm) 4.02 (s, 3H), 7.08 (m, 1 Hz), 7.17 (dt, 3 1H, J ) 5.4, 8.2 Hz), 7.40 (d, 1H, J ) 7.9 Hz), 8.28 (s, 1H). 13C NMR (CDCl3, δ, ppm) 62.5 (CH3), 115.4 (d, CH, J ) 22.2 Hz), 120.1 (d, C, J ) 14.5 Hz), 123.9 (C), 128.6 (CH), 130.6 (d, CH, J ) 9.4 Hz), 144.0 (CH), 160.1 (d, C, J ) 257 Hz). For Z-isomer 35a: 1H NMR (CDCl3, δ, ppm) 3.95 (s, 3H), 7.05 (m, 1H), 7.20 (dt, 1H, J ) 5.8, 7.9 Hz), 7.36 (d, 1H, J ) 7.9 Hz), 8.28 (s, 1H). 13C NMR (CDCl , δ, ppm) 62.5 (CH ), 114.4 (d, CH, J ) 22 Hz), 3 3 121.8 (d, C, J ) 14.5 Hz), 122.1 (C), 127.8 (CH), 130.9 (d, CH, J ) 9.0 Hz), 140.0 (CH), 159.0 (d, C, J ) 254 Hz). 2-Chloro-6-fluorobenzaldehyde O-Methyloxime (36). The oxime solution was prepared according to the method B. It was concentrated in vacuo, and the residue was chromatographed on silica gel (eluent 10:1 heptane-ethyl acetate) to give pure oxime 36 as a 22:78 mixture of Z/E isomers as determined by 1H NMR.
8172 J. Org. Chem., Vol. 71, No. 21, 2006
For the E-isomer (36b): 1H NMR (CDCl3, δ, ppm) 4.02 (s, 3H), 6.93-7.09 (m, 1H), 7.17-7.31 (m, 2H), 8.32 (s, 1H). 13C NMR (CDCl3, δ, ppm) 62.5 (CH3), 114.7 (d, CH, J ) 23 Hz), 118.7 (d, C, J ) 14 Hz), 125.4 (CH), 130.2 (d, CH, J ) 9.0 Hz), 134.5 (C), 142.1 (CH), 160.3 (d, C, J ) 257 Hz). Anal. Calcd for C8H7ClFNO: C, 51.22; H, 3.76; N, 7.47. Found: C, 51.40; H, 3.86; N, 7.30. 2-Chloro-6-hydrazidobenzonitrile (38). Hydrazine (98%, 5 mL) was added over 15 min to a solution of 2-chloro-6-fluorobenzonitrile (0.78 g, 5 mmol) in THF (5 mL) while maintaining the internal temperature below 30 °C. After an additional 0.5 h at room temperature the THF layer was separated and concentrated to dryness in vacuo maintaining the internal temperature below 30 °C to give hydrazide 38 (0.75 g, 90%): 1H NMR (DMSO-d6, δ, ppm) 4.37 (s, 2 H), 6.73 (d, 1H, J ) 7.7 Hz), 7.17 (d, 1H, J ) 8.6 Hz), 7.38 (t, 1H, J ) 8.3 Hz), 7.64 (s, 1H). 13C NMR (DMSO-d6, δ, ppm) 91.9 (C), 110.2 (CH), 114.76 (C), 115.6 (CH), 134.0 (CH), 134.7 (C), 155. 1 (C). Anal. Calcd for C7H6ClN3: C, 50.17; H, 3.61; N, 25.07. Found: C, 49.78; H, 3.54; N, 24.97. Heating of the above reaction mixture to 75 °C for 1 h resulted in the quantitative conversion of intermediate 38 into aminoindazole 34b, as determined by HPLC methodology. JO0613784
